Results are presented from a consecutive series of 396 patients referred to hospital in one of the areas of highest incidence in the United Kingdom in a unit that saw all referred cases for assessment and subsequent decision on the most appropriate therapeutic modality. Standard criteria were applied when selecting those patients in whom an attempt at curative resection was justified. Of the 396 patients, only 164 (41%) were considered fit enough to undergo thoracoabdominal resection and to have a tumor that had not invaded contiguous structures or metastasized. The remainder were treated by palliative intubation and more recently by laser photocoagulation. Of the patients who underwent resection, most had an Ivor Lewis esophagogastrectomy and lymphadenectomy, those with proximal lesions undergoing three-stage total esophagectomy. The overall 30-day mortality was 8.3%, falling to 6.7% during the period 1985-1990 when thoracic epidural analgesia was used routinely and fatal respiratory complications were avoided. Overall 5-year survival rate was 23%. However, stratification for tumor staging showed that the 5-year survival for node-negative patients was 47% and for tumors confined within the adventitia 71%. Results of surgical treatment of squamous esophageal carcinoma in the West are considerably more depressing than those in many Eastern series. This difference is largely related to the late presentation in the West, as the relatively low incidence does not justify screening programs. However, results equivalent to those in Eastern series are obtained for the small subset of patients who do present with early disease.(ABSTRACT TRUNCATED AT 250 WORDS)